New synthesis, electroluminescence, and photophysical properties of poly[(formylphenyl)methylsilanediyl] and its derivatives

Article abstract of DOI:10.1021/ma010736w

A new one-step synthesis of formylated poly[methyl(phenyl)silanediyl] (PMPSi) is reported. The aldehyde groups were incorporated into the parent polymer by the reaction with dichloromethyl methyl ether in the presence of Lewis acid (SnCl₄). The

Full text of DOI:10.1021/ma010736w

Macromolecules 2002, 35, 3463-3473
3463
New Synthesis, Electroluminescence, and Photophysical Properties of
Poly[(formylphenyl)methylsilanediyl] and Its Derivatives
Dr a h om ı´r Vy´p r a ch tick y´* a n d Veˇr a Cim r ova´
Institute of Macromolecular Chemistry, Academy of Sciences of the Czech Republic, Heyrovsky´ Sq. 2,
162 06 Prague 6, Czech Republic
Received April 30, 2001; Revised Manuscript Received February 4, 2002
ABSTRACT: A new one-step synthesis of formylated poly[methyl(phenyl)silanediyl] (PMPSi) is reported.
The aldehyde groups were incorporated into the parent polymer by the reaction with dichloromethyl
methyl ether in the presence of Lewis acid (SnCl4). The reaction takes place via an unstable aryl(methoxy)-
methyl chloride intermediate. The new procedure is an important simplification of the known four-step
synthesis of the poly[(formylphenyl)methylsilanediyl] via chloromethylation in the first step. The presence
and reactivity of aldehyde groups in the polymer were proved by means of 1H NMR, UV-vis, and FT IR
spectroscopies. The reduction in relative molecular weight of the parent polysilane after reaction was
not so pronounced as in the multistep synthesis. Model compounds N-benzylidene-1-aminopyrene,
N-benzylidene-4-phenylaniline, and N-benzylidene-4-phenylazoaniline were synthesized for spectral
characterization of the substituted polysilanes. The π-conjugated systems (pyrene, biphenyl, azobenzene)
were attached to the σ-conjugated polysilane backbone. Photophysical properties, photostability, charge
photogeneration, and photoluminescence of these materials were studied. Single- and double-layer light-
emitting devices (LEDs) with emission in visible spectral region were fabricated. Their spectral and
electrical characteristics were studied. Electroluminescence efficiency in double-layer LEDs reached the
value of 0.03-0.06% and was more than 1 order of magnitude higher than that in single-layer devices
(∼0.005%). Modified polysilanes exhibited better photostability than PMPSi and are particularly intended
for the use in multilayer or blend devices.
In tr od u ction
(phenyl)silanediyl] (PMPSi) and chloromethyl methyl
ether generated in situ from dimethoxymethane and
thionyl chloride. The chloromethyl group can undergo
a host of reactions. The transformation of chloromethyl
into aldehyde groups enabled preparation of polysi-
lanes15 bearing directly attached π-conjugated chro-
mophores, which can be useful in many applications.
Four reaction steps, however, were necessary to ac-
complish the synthesis of poly[(formylphenyl)methyl-
silanediyl] from the parent PMPSi.
Our previous studies of charge transport and photo-
generation in substituted PMPSi showed that these
processes were strongly influenced by the nature of
substituents. We have shown16 that aromatic ring
substitution with an electron-withdrawing group in-
creased the quantum efficiency of charge photogenera-
tion by a factor higher than 1 order of magnitude. The
substitution of PMPSi also influenced the hole mobility,
which depends on electron affinities and dipole moments
of the attached chromophores. In contrast to the charge
photogeneration and charge transport data, there is no
information concerning electroluminescence in such
PMPSi derivatives. PMPSi was used as a hole-trans-
porting layer in multilayer LEDs.17,18 PMPSi,19,20
dialkyl,21-24 diaryl,23,25-27 silole-incorporated,28 and oxy-
gen-cross-linked29 polysilanes, and PMPSi copolymers
containing thiophene30,31 or anthracene32 units in the
polymer backbone were reported as emitting layers in
polymer LEDs. Recently,33 we have shown an improved
performance of polymer LEDs based on blends composed
of poly[2,5-bis(isopentyloxy)-1,4-phenylene] (SPPP) and
PMPSi compared to the single-component LEDs made
of SPPP.
In the past years polysilanes have attracted consider-
able attention as a new class of σ-conjugated polymers.1,2
Although the synthesis of high-molecular-weight soluble
polysilanes was reported nearly 20 years ago,3-5 the
functionalization or modification of such polymers is still
a subject of investigation. The problem is that not all
reactive functional groups present in the dichlorosilane
monomer (R1R2SiCl2; R1, R2 ) alkyl or aryl) before
polymerization can withstand drastic conditions of the
Wurtz coupling reaction. Although the toluene-sodium
dispersion reflux conditions are often replaced by reduc-
tive coupling at lower temperatures (using crown ethers6
or graphite-potassium7 as a phase-transfer or hetero-
geneous catalyst, respectively), the main problem still
remains. Hence, the direct syntheses of functionalized
polysilanes have been restricted to dichlorosilane mono-
mers with functional groups, which are resistant to the
reaction conditions (aryloxy,1,8 dimethylamino,8,9 and
trimethylsilyl10 groups). The trimethylsilyl group has
been used to protect the hydroxy group in dichloro[1-
(4-hydroxy-3-methoxyphenyl)propyl]phenylsilane prior
to polymerization.11 Chemical modification of the poly-
mer affords another possibility of introducing a reactive
functional group into parent polysilanes. However, the
choice of suitable reactions is limited because of a low
oxidation potential of the σ-conjugated silicon backbone.1
Chlorine- or bromine-containing polymers were pre-
pared by addition of HCl or HBr in the presence of Lewis
acid to silane copolymers carrying 2-(cyclohex-3-en-1-
yl)ethyl substituents.12 A very useful chemical modifica-
tion of aryl-containing polysilanes is the Friedel-Crafts
chloromethylation13 using chloromethyl methyl ether
and stannic chloride as a catalyst. The reaction was
recently studied in more detail14 with poly[methyl-
In this paper we developed a one-step synthesis of
formylated PMPSi, which was further used as a func-
tionalized polymeric precursor for the attachment of
10.1021/ma010736w CCC: $22.00 © 2002 American Chemical Society
Published on Web 03/28/2002

3464 Vy´prachticky´ and Cimrova´
Macromolecules, Vol. 35, No. 9, 2002
(THF): λmax ) 331 nm, ꢀ ) 16 700 L mol-1 cm-1 (Figure 3).
FT IR (film on KBr): 3055, 3030 (aromatic C-H); 1624 (CdN);
1169 (aryl-N, stretching); 970 (azomethine C-H, in-plane);
880 (azomethine C-H, out-of-plane); 838, 760, 721, and 689
cm-1 (aromatic C-H, out-of-plane).
Syn th eses of P olym er s. P oly[m eth yl(ph en yl)silan ediyl].
PMPSi-1. The polymer was prepared by Wurtz coupling
polymerization of dichloromethyl(phenyl)silane. A sulfonation
flask (1500 mL) with reflux condenser, pressure equalizing
addition funnel, and a motor-driven glass stirrer was charged
under argon with sodium dispersion (33 g, 1.43 mol of Na) in
toluene (400 mL) excluding light from the reaction mixture.
The stirred sodium dispersion was heated to a gentle reflux
and a solution of dichloromethyl(phenyl)silane (120 g, 0.63 mol)
in toluene (100 mL) was added quickly enough to maintain
the reflux. The reaction mixture was refluxed for 3 h, then
cooled to room temperature, and quenched with ethanol (150
mL) and water (700 mL). The toluene layer was separated and
washed with water (2 × 750 mL), and the solvent was
evaporated. The greasy residue was dissolved in tetrahydro-
furan (THF, 250 mL) and the addition of a propan-2-ol-
methanol mixture (2 L, 1:1 by vol) precipitated the PMPSi as
a white powder; yield 51.9 g (68%). The oligomers were further
removed by extraction with boiling diethyl ether; yield 29.6 g
(39%). The purified polymer possessed a largely unimodal but
very broad molar mass distribution (Mw ) 39 900, Mn ) 8600).
1H NMR (CDCl3): -1.1 to 0.2 (broad peak, 3H, CH3), 6.1-7.3
(broad peak, 5H, C6H5) ppm. UV-vis (THF): λmax ) 338 nm,
ꢀ ) 8200 L mol-1 cm-1 (characteristic σ f σ* transition for
the silicon backbone of PMPSi). FT-IR (film on KBr): 3070,
3050, and 3020 (aromatic C-H); 2960 and 2895 (aliphatic
C-H); 1950, 1890, and 1815 (aromatic); 1485 (skeletal ring
breathing); 1430 and 1100 (Si-Ph); 1250 (Si-Me); 730 (aro-
matic C-H, out-of-plane); 696 (Si-C); 462 cm-1 (Si-Si).
PMPSi-2. Poly[methyl(phenyl)silanediyl] was also prepared
using sodium dispersion in toluene at 65 °C and 18-crown-6
as a phase-transfer catalyst. Addition of the crown ether to
the heterogeneous Wurtz reaction promotes the synthesis of
a high-molecular-weight polymer.6 Sodium dispersion (3.9 g,
0.17 mol of Na) and 18-crown-6 (0.012 g, 4.54 × 10-5 mol) were
stirred under argon in dry toluene (75 mL) at 65 °C. Dichloro-
methyl(phenyl)silane (14.09 g, 0.074 mol) in dry toluene (75
mL) was added dropwise (30 min), and the reaction mixture
was heated at 65 °C for 3 h with strong mixing. After cooling,
the reaction was quenched with methanol (50 mL), and toluene
(100 mL) and water (100 mL) were added. The toluene layer
was washed with water (3 × 150 mL) and dried over anhy-
drous magnesium sulfate (2 h). The solvent was evaporated;
the residue was dissolved in THF (150 mL) and precipitated
into propan-2-ol (900 mL); yield 2.82 g (32%). The polymer
showed a bimodal molar mass distribution. The low-molecular-
weight fraction was removed by stirring (2 h) with hexane (50
mL). The high-molecular-weight fraction was filtered and
reprecipitated from THF into propan-2-ol; yield 1.09 g (12%).
Mw ) 321 000, Mn ) 205 000. 1H NMR and FT IR spectra
confirmed the structure of PMPSi (see above). UV-vis
F igu r e 1. Structures of the polymers under study.
π-conjugated systems (Figure 1). Single-layer and double-
layer light-emitting devices (LEDs) were fabricated, and
their electric and optical characteristics were measured
and discussed.
Exp er im en ta l Section
Ma ter ia ls. Sodium (∼30% dispersion of Na in toluene),
dichloro(methyl)phenylsilane, 1-aminopyrene, 4-aminoazoben-
zene, 4-aminobiphenyl, 4-methylbenzene-1-sulfonic acid, benz-
aldehyde, 1,4,7,10,13,16-hexaoxacyclooctadecane (18-crown-6),
and dichloromethyl methyl ether were the commercial prod-
ucts (Fluka, Aldrich) and were used as received. Toluene p.a.
(Lachema Brno, Czech Republic) was dried by 24 h refluxing
over sodium metal and distilled.
Syn th eses of Model Com pou n ds. N-Benzylidene-1-amino-
pyrene (NB-Pyr). 1.30 g (0.006 mol) of 1-aminopyrene, 5.31 g
(0.05 mol) of benzaldehyde, and 0.05 g of 4-methylbenzene-1-
sulfonic acid (catalyst) were dissolved in 80 mL of toluene. The
mixture was refluxed 5 h using a Dean-Stark adapter for
water removing. After cooling the toluene was evaporated and
methanol (50 mL) was added. The yellow crystals were filtered
off, recrystallized from ethanol, and dried. Yield 1.35 g (74%);
mp 99-100 °C. Anal. Calcd for C23H15N (305.39): C, 90.46;
H, 4.95; N, 4.59%. Found: C, 90.41; H, 5.11; N, 4.56%. 1H NMR
(THF-d8): 7.47 (t, 5H, C6H5), 7.75-8.16 (m, 8H, C16H9), 8.65
(d, 1H, C16H9), 8.76 (s, 1H, NdCH) ppm (Figure 2). UV-vis
(THF): λmax ) 382 nm, ꢀ ) 24 300 L mol-1 cm-1 (Figure 3).
FT-IR (film on KBr): 3040 (aromatic C-H); 1620 (CdN); 1582,
1453 (pyrene ring stretching); 1180 (aryl-N, stretching); 1100
(aromatic C-H, bending); 970 (azomethine C-H, in-plane);
840 (pyrene ring deformation); 757, 719, and 688 cm-1
(aromatic C-H, out-of-plane).
N-Benzylidene-4-phenylazoaniline (NB-Azo) was prepared
by the same procedure as NB-Pyr above, using 4-aminoazo-
benzene (3.0 g, 0.015 mol), benzaldehyde (10.62 g, 0.1 mol),
and 4-methylbenzene-1-sulfonic acid (0.1 g) as a reaction
mixture. Yield 3.47 g (80%); mp 130 °C (lit.34 130-131 °C).
Anal. Calcd for C19H15N3 (285.34): C, 79.98; H, 5.30; N, 14.73%.
Found: C, 79.95; H, 5.38; N, 14.69%. 1H NMR (THF-d8): 7.29-
7.45 (m, 8H, i.e., 5H of C6H5 and 3H of C6H5-N), 7.83-7.93
(m, 6H, i.e., 2H of C6H5-N and 4H of N-C6H4-N), 8.54 ppm
(s, 1H, NdCH) (Figure 2). UV-vis (THF): λmax ) 360 nm,
ꢀ ) 29 200 L mol-1 cm-1 (Figure 3). FT IR (film on KBr):
3055, 3025 (aromatic C-H); 1621 (CdN); 1445 (NdN); 1187
(aryl-N, stretching); 967 (azomethine C-H, in-plane); 875
(azomethine C-H, out-of-plane); 850, 754, and 685 cm-1
(aromatic C-H, out-of-plane).
N-Benzylidene-4-phenylaniline (NB-BiPh) was prepared by
the same procedure as NB-Pyr above, using 4-aminobiphenyl
(3.05 g, 0.018 mol), benzaldehyde (10.62 g, 0.1 mol), and
4-methylbenzene-1-sulfonic acid (0.1 g) as a reaction mixture.
Yield 3.01 g (65%); mp 148 °C (lit.35 147 °C, lit.36 148-149 °C).
Anal. Calcd for C19H15N (257.34): C, 88.68; H, 5.88; N, 5.44%.
Found: C, 88.59; H, 6.03; N, 5.43%. 1H NMR (THF-d8): 7.21-
7.40 (m, 8H, i.e., 5H of C6H5 and 3H of C6H5-Ph), 7.55-7.60
(m, 4H, i.e., 2H of C6H5-Ph and 2H of Ph-C6H4-N), 7.84-
7.90 (m, 2H, Ph-C6H4-N), 8.51 ppm (s, 1H, NdCH). UV-vis
(THF): λmax ) 338 nm, ꢀ ) 9200 L mol-1 cm-1
.
P oly[(for m ylp h en yl)m et h ylsila n ed iyl]. PMPSi-CHO-
(12 mol %). Under argon and with exclusion of light, PMPSi-1
(4.81 g, 0.04 mol) was dissolved in dichloromethane (100 mL)
and cooled to 0 °C. SnCl4 (20.84 g, 0.08 mol) and dichloro-
methyl methyl ether (9.20 g, 0.08 mol) were added, and the
mixture was stirred at 0 °C for 2 h and then poured onto
crushed ice (∼150 g). The organic layer was separated and
washed with water (2 × 100 mL). The solvent was evaporated;
the residue was dissolved in THF (80 mL) and precipitated
into methanol (700 mL). The white polymer was filtered off
and dried; yield 2.90 g. Mw ) 21 000, Mn ) 7500. 1H NMR
(CDCl3): -1.1 to 0.6 (broad peak, 3H, CH3), 6.1-7.6 (broad
peak, 5H, C6H5), 9.5-9.9 ppm (peak, 1H, CHO) (Figure 4).
From 1H NMR spectrum the content of aldehyde groups in
PMPSi-CHO was determined as 12 mol %. UV-vis (THF):
λmax ) 330 nm, ꢀ ) 6500 L mol-1 cm-1. FT IR (film on KBr):
vibration bands of PMPSi and 2926 (aldehyde C-H), 1702
cm-1 (aldehyde CdO).

Macromolecules, Vol. 35, No. 9, 2002
PMPSi and Its Derivatives 3465
F igu r e 2. 1H NMR spectra of model compounds (NB-Pyr, NB-Azo) and corresponding modified polymers PMPSi-Pyr(12 mol
%) and PMPSi-Azo(12 mol %) in deuterated tetrahydrofuran.
F igu r e 4. 1H NMR spectrum of PMPSi-CHO(12 mol %) in
CDCl3. The sharp signal at 0 ppm belongs to the internal
standard (hexamethyldisiloxane). The numbers denote the
integrated intensities for aldehyde (1H, CHO, 1.184) and
aromatic (5H, C6H5, 53.867) protons.
Various reaction conditions were studied, and the results
are summarized in Table 1.
F igu r e 3. UV-vis spectra of model compounds NB-Azo
(dashed line), NB-Pyr (solid line), and NB-BiPh (dashed and
dotted line) in tetrahydrofuran.
Mod ifica t ion of P MP Si-CH O w it h 1-Am in op yr en e.
PMPSi-Pyr(12 mol %). 0.71 g (0.7 mmol of -CHO groups) of
PMPSi-CHO (12 mol %) was refluxed with 0.76 g (3.5 mmol)
of 1-aminopyrene and 0.05 g of 4-methylbenzene-1-sulfonic
acid in toluene (60 mL) for 30 min. After cooling, the reaction
mixture was precipitated into methanol (500 mL). The bright
yellow polymer was filtered off and dried. The raw material
was three times reprecipitated from THF (30 mL) into
PMPSi-CHO(6 mol %). Under argon and with exclusion of
light, PMPSi-1 (4.81 g, 0.04 mol) was dissolved in dichloro-
methane (100 mL) and cooled to 0 °C. SnCl4 (15.63 g, 0.06 mol)
and dichloromethyl methyl ether (6.90 g, 0.06 mol) were added,
and the mixture was stirred at 0 °C for 30 min and then poured
onto crushed ice (∼150 g). Then the procedure was the same
methanol (500 mL); yield 0.65 g (76%). Mw ) 23 000, Mn
)
1
1
as above; yield 3.45 g. Mw ) 33 000, Mn ) 8200. H NMR and
8100. H NMR (THF-d8): -1.1 to 0.6 (broad peak, 3H, CH3),
6.1-7.6 (broad peak, 5H, C6H5), 7.6-8.25 (m, 9H, C16H9), 8.4-
8.8 (peak, 1H, NdCH), 9.5-9.9 (traces, 1H, CHO) ppm (Figure
2). UV-vis (THF): λmax ) 330 nm (σ f σ*), ꢀ330 ) 7600 L mol-1
FT IR spectra confirmed the presence of aldehyde groups, and
their content was determined as 6 mol %. UV-vis (THF): λmax
) 335 nm, ꢀ ) 7300 L mol-1 cm-1
.

3466 Vy´prachticky´ and Cimrova´
Macromolecules, Vol. 35, No. 9, 2002
Ta ble 1. Rea ction Con d ition s for F or m yla tion of P MP Si-1 w ith Dich lor om eth yl Meth yl Eth er (Cl2CHOCH3) in th e
P r esen ce of Lew is Acid s (Sn Cl4, TiCl4) a t 0 °C (Rela ted p er 1 m ol of P MP Si-1)
UV-vis content (mol %)
expt
no.
SnCl4
(mol)
TiCl4
(mol)
Cl2CHOCH3
(mol)
time
(min)
1H NMR CHO
content (mol %)
PMPSi-Pyr
(382 nm)
PMPSi-Azo
(360 nm)
1
2
3
4a
5
6
7
8
0.5
1.5
2.0
2.0
2.0
0.5
1.5
2.0
2.0
2.0
0.8
0.4
1.5
10
30
120
270
480
30
0
6
12
10
10
b
<0.5
6
12
10
11
b
<0.5
5
11
9
10
b
b
1.7
0.9
0.5
30
20
b
1
b
1
1
a
b
At 20 °C. The aldehyde content was not determined because oily oligomeric products were formed.
C-C6H4-Si), 7.35-7.5 (peak, 3H, C6H5-N), 7.75-8.0 (peak,
2H of C6H5-N and 4H of N-C6H4-N), 8.2-8.5 (peak, 1H,
NdCH), 9.5-9.9 (traces, 1H, CHO) ppm (Figure 2). UV-vis
(THF): λmax ) 335 nm (σ f σ*), ꢀ335 ) 7800 L mol-1 cm-1; λ )
360 nm (shoulder, π f π*, n f π*), ꢀ360 ∼ 3000 L mol-1 cm-1
(Figure 5). FT IR (film on KBr): vibration bands of PMPSi
and 1620 (CdN), 1444 (NdN), 1182 (aryl-N, stretching), 756
cm-1 (aromatic C-H, out-of-plane).
PMPSi-Azo(6 mol %). The same procedure as above but with
0.92 g (0.45 mmol of -CHO groups) of PMPSi-CHO(6 mol
%), 1.50 g (7.6 mmol) of 4-aminoazobenzene, and 0.1 g of
4-methylbenzene-1-sulfonic acid as starting materials; yield
0.58 g (59%). Mw ) 34 000, Mn ) 8500. UV-vis (THF): λmax
)
335 nm (σ f σ*), ꢀ335 ) 7800 L mol-1 cm-1; λ ) 360 nm
(shoulder, π f π*, n f π*), ꢀ360 ∼ 1500 L mol-1 cm-1 (Figure
5). 1H NMR and FT IR data are identical with those for PMPSi-
Azo(12 mol %), but the peaks responsible for azobenzene
substitution are reduced in their intensity approximately to
one-half.
Mod ifica tion of P MP Si-CHO w ith 4-Am in obip h en yl.
PMPSi-BiPh(3 mol %). 0.71 g (0.7 mmol of -CHO groups) of
PMPSi-CHO (12 mol %) was refluxed with 1.20 g (7 mmol) of
4-aminobiphenyl and 0.1 g of 4-methylbenzene-1-sulfonic acid
in toluene (100 mL) for 30 min. After cooling, the reaction
mixture was precipitated into methanol (700 mL). The white
polymer was filtered off and dried. The raw material was three
times reprecipitated from THF (50 mL) into methanol (700
mL); yield 0.41 g (51%). Mw ) 20 000, Mn ) 7900. 1H NMR
(THF-d8): -1.1 to 0.4 (broad peak, 3H, CH3), 6.1-7.25 (broad
peak, 5H of C6H5-Si and 3H of C6H5-Ph), 7.25-7.5 (peak,
2H of C6H5-Ph and 2H of Ph-C6H4-N), 7.55-7.75 (peak, 2H,
Ph-C6H4-N), 8.2-8.4 (broad peak, 1H, NdCH), 9.5-9.9 ppm
(broad peak, 1H, CHO). The ratio of the integrated intensities
of aldehyde (CHO) and azomethine (NdCH) hydrogens was
found to be 3:1. It means that a quarter of aldehyde groups of
PMPSi-CHO(12 mmol %) was reacted with 4-aminobiphenyl
to give PMPSi-BiPh(3 mol %). UV-vis (THF): λmax ) 335 nm
(σ f σ*), ꢀ335 ) 8100 L mol-1 cm-1; shoulder at λ ) 360-400
nm. FT IR (film on KBr): vibration bands of PMPSi and 1699
(aldehyde CdO), 1620 (CdN), 1170 (aryl-N, stretching), 756
cm-1 (aromatic C-H, out-of-plane).
F igu r e 5. UV-vis spectra of PMPSi-1 (solid line) and
polymers modified with pyrene structural units, PMPSi-Pyr-
(6 mol %) (dashed and dotted line) and PMPSi-Pyr(12 mol
%) (dashed double dotted line), and polymers modified with
azobenzene structural units, PMPSi-Azo(6 mol %) (dotted
line) and PMPSi-Azo(12 mol %) (dashed line), in tetrahydro-
furan.
cm-1; λmax ) 384 nm (π f π*, n f π*), ꢀ384 ) 3020 L mol-1
cm-1 (Figure 5). FT IR (film on KBr): vibration bands of
PMPSi and 1620 (CdN), 1589 (pyrene ring stretching), 1183
(aryl-N, stretching), 841 (pyrene ring deformation), 756 cm-1
(aromatic C-H, out-of-plane).
PMPSi-Pyr(6 mol %). The same procedure as above but
with 0.71 g (0.35 mmol of -CHO groups) of PMPSi-CHO (6
mol %) and 0.39 g (1.8 mmol) of 1-aminopyrene as starting
materials; yield 0.58 g (74%). Mw ) 30 000, Mn ) 8100. UV-
vis (THF): λmax ) 335 nm (σ f σ*), ꢀ335 ) 7800 L mol-1 cm-1
;
P olym er Ch a r a cter iza tion . Molecular weights of the poly-
mers were determined using GPC/SEC chromatograph (Labo-
ratory Instruments, CZ) equipped with RI and UV 254 nm
detectors, column 8 × 600 mm (PSS 10 000, PSS, Germany)
and data collection and treatment (Data Monitor, Watrex, CZ).
Polystyrene standards (PL Laboratories, UK) were used for
calibration. The samples were measured in THF, which was
dried over molecular sieve 4A and distilled.
λmax ) 384 nm (π f π*, n f π*), ꢀ384 ) 1490 L mol-1 cm-1
1
(Figure 5). H NMR and FT IR data are identical with those
for PMPSi-Pyr(12 mol %), but the peaks responsible for
pyrene substitution are reduced in their intensity approxi-
mately to one-half.
Mod ifica tion of P MP Si-CHO w ith 4-Am in oa zoben -
zen e. PMPSi-Azo(12 mol %). 0.92 g (0.9 mmol of -CHO
groups) of PMPSi-CHO(12 mol %) was refluxed with 3.0 g
(15.2 mmol) of 4-aminoazobenzene and 0.1 g of 4-methylben-
zene-1-sulfonic acid in toluene (120 mL) for 30 min. After
cooling, the reaction mixture was precipitated into methanol
(800 mL). The yellow polymer was filtered off and dried. The
raw material was three times reprecipitated from THF (50 mL)
into methanol (700 mL); yield 0.71 g (66%). Mw ) 22 500, Mn
) 7800. 1H NMR (THF-d8): -1.0 to 0.5 (broad peak, 3H, CH3),
6.1-7.2 (broad peak, 5H, C6H5-Si), 7.2-7.35 (peak, 4H,
UV-vis spectra were taken on a Perkin-Elmer Lambda 20
spectrometer. Both solutions and films on fused silica sub-
strates were measured. The contents of azobenzene or pyrene
structural units in polysilanes (Table 1) were determined by
spectrometry in THF using molar absorption coefficients of
model compounds NB-Azo (ꢀ360 ) 29 200 L mol-1 cm-1) or
NB-Pyr (ꢀ382 ) 24 300 L mol-1 cm-1).
1H NMR spectra were taken on a Bruker ACF-300 spec-
trometer at 300.1 MHz in CDCl3 or deuterated THF using

Macromolecules, Vol. 35, No. 9, 2002
PMPSi and Its Derivatives 3467
hexamethyldisiloxane as an internal standard. The content of
aldehyde groups in PMPSi-CHO was determined from the
integrated intensity of the CHO signal at 9.5-9.9 ppm (1H)
by comparison with a broad signal of aromatic protons at 6.1-
7.6 ppm (5H).
FT IR spectra were measured using a Perkin-Elmer Paragon
1000 PC Fourier transform infrared spectrometer in KBr
pellets or as a film on KBr pellet.
from the photocurrent Iph caused by the emitted photons using
the relation
Iph
nph
)
(1)
hc
n(λ) S(λ) dλ
∫
λ
where λ is the wavelength of the emitted photons, n(λ) the
fraction of photons with wavelength λ, hc/ λ the photon energy,
and S(λ) the sensitivity of the photodiode (A W-1). Iph was
evaluated from the measured voltage signal U using Iph ) U/G,
where G (V A-1) is the gain of the built-in amplifier. n(λ) was
calculated from the EL spectrum. For the estimation of the
internal EL quantum efficiency, ηEL, the external EL efficiency
was multiplied by the factor 2n2, where n is the refractive index
of the polymer.37 The value n ) 1.67 was used in our evaluation
Sa m p le P r ep a r a tion . Thin polymer films were prepared
by spin-coating from toluene solutions. All solutions were
filtered with 0.45 µm Millex-FH13 Millipore syringe filters prior
the spin-coating. Thin films for optical studies were spin-coated
onto fused silica substrates. All films exhibited a good optical
quality. Polymer LEDs with a hole-injecting indium tin oxide
(ITO) electrode and an electron-injecting aluminum electrode
were fabricated. Two types of the devices were prepared: the
polymer layers were spin-coated (1) onto ITO-covered glass
substrates and (2) onto ITO substrates covered with thin layer
of poly[3,4-(ethylenedioxy)thiophene]/poly(styrenesulfonate)
(PEDT:PSS). The ITO glass substrates were purchased from
Merk (Germany) and PEDT:PSS Baytron-P from Bayer AG
(Germany). The 50 nm thick PEDT:PSS layers were prepared
by spin-coating and dried in a vacuum at 396 K for 3 h prior
the polymer film deposition. Finally, 60-80 nm thick top
aluminum (Al) electrodes were vacuum-evaporated on the top
of polymer films to form LED devices. Typical active areas of
the LEDs were 4 or 8 mm2. Samples for charge photogenera-
tion measurements (film thicknesses from 700 to 1300 nm)
were prepared by spin-coating of the polymer solution in
toluene onto stainless steel substrates. All polymer films were
dried in a vacuum (10-3 Pa) at 323 K for 6 h. Layer thicknesses
were measured using a KLA-Tencor P-10 profilometer. All
thin-film preparation and the LED fabrication were done
under ambient laboratory conditions.
of the ηEL
.
Ch a r ge P h otogen er a tion Mea su r em en ts. The measure-
ments of the photogeneration efficiencies were carried out
using the technique of emission-limited photoinduced dis-
charge.38 A sample was charged to the initial surface potential
U0 ) Q0/C (Q0 is the surface charge density and C is the sample
capacitance per unit area) by a corona and then discharged
upon irradiation. The photoinduced discharge quantum ef-
ficiency η′ was calculated from the formula
η′ ) -(1/ eΦ)(dQ/ dt)0 ) -(1/eΦ)C(dU/ dt)0
(2)
where Φ is the absorbed photon flux density, (dQ/dt)0 the rate
of change of the surface charge density, and (dU/dt)0 the initial
photoinduced discharge slope. Under emission-limited condi-
tions, the η′ is independent both of the sample thickness and
of the light intensity and is equal to the photogeneration
efficiency ηph. The surface potentials were measured with a
homemade rotary electrodynamic condenser electrometer. The
measurements were performed with irradiation incident upon
the positively charged surface, using a 300 W xenon lamp
(Oriel) with a band filter maximum at 358 nm (a bandwidth
of 65 nm). To ensure emission-limited conditions, the discharge
Cyclovolta m m etr ic Mea su r em en ts. Cyclic voltammetry
(CV) was measured using a PA4 polarographic analyzer
(Laboratory Instruments, CZ) with a tree-electrode cell in
dichloromethane containing tert-butylammonium perchlorate
(0.1 mol L-1) as an electrolyte. Platinum strip electrodes were
used as the working and counter electrodes, and Ag/AgCl
electrode was used as the reference electrode. Polymer con-
centration in the measured solution was 10-3 mol L-1. Typical
was measured with photon fluxes as low as possible (1017
-
1018 photons m-2 s-1). Light-induced discharge rates were
corrected for the dark discharge rate. The sample capacitances
C were determined from capacitance measurements at the
frequency 1 kHz.
scan rates were 20-100 mV s-1
.
Electr olu m in escen ce a n d P h otolu m in escen ce Mea -
su r em en ts. Electroluminescence (EL) and photoluminescence
(PL) spectra were measured using a homemade spectrofluo-
rometer with single photon-counting detection (SPEX, RCA
C31034 photomultiplier). LEDs were supplied from a Keithley
237 source measure unit, which served for the simultaneous
recording of the current flowing trough the sample. PL spectra
were taken perpendicular to the sample surface, and the angle
of incidence for the excitation beam was 30° relative to the
sample surface normal to minimize reabsorption. A 300 W
xenon lamp (Oriel) was used as the excitation source. The
measured EL and PL emission spectra were corrected for the
spectral response of the detection system. Current-voltage and
luminance-voltage characteristics were recorded simulta-
neously using the Keithley 237 source measure unit and a
silicon photodiode with integrated amplifier (EG&G HUV-
4000B) for the detection of total light output. A voltage signal
from the photodiode was recorded with Hewlett-Packard
34401A multimeter. The LED characteristics were measured
in a vacuum chamber to prevent any electrode degradation.
Photostability was followed by UV-vis spectroscopy. PL and
absorption measurements were performed under ambient
laboratory conditions.
Resu lts a n d Discu ssion
Syn th eses a n d Mod ifica tion of P MP Si-CHO.
The modification of PMPSi by the attachment of π-con-
jugated chromophores (Figure 1) drove the effort to
simplify the known four-step synthesis of poly[(formyl-
phenyl)methylsilanediyl].15 The previous authors used
the Kroehnke method for the synthesis of chemically
sensitive aldehydes, where the pyridinium salt of the
chloromethylated arene was reacted with a nitroso
compound yielding the corresponding nitrone, which
could be hydrolyzed to the aldehyde. The attempts to
convert chloromethylated PMPSi directly to the alde-
hyde using either high-temperature oxidation with
dimethyl sulfoxide39 or the Sommelet reaction10 were
not successful. Vilsmeier-Haack, Gattermann-Koch,
and Gattermann reactions are relatively simple and the
most frequently used procedures for formylation of
aromatic hydrocarbons. Unfortunately, the catalysts
(AlCl3, ZnCl2, POCl3) employed are too aggressive to
preserve the silicon backbone; the less aggressive TiCl4,
SnCl4, and SbCl4 were found to be inefficient.40
The EL quantum efficiencies were evaluated from the
measurements of EL spectra and the total light output. The
external EL efficiency is expressed by the relation ηext ) nph
/
nel, where nph and nel are the total number of photons emitted
from LED and the total number of electrons injected into the
polymer per second, respectively. nel was evaluated from the
current I passing trough LED as nel ) I/ e , where e ) 1.6 ×
10-19 C. The total light output was measured with calibrated
silicon photodiode with built-in amplifier. nph was determined
On the other hand, in the synthesis of aromatic
aldehydes using dichloromethyl alkyl ethers, the cata-
lysts AlCl3, TiCl4, and SnCl4 have been used success-
fully. One-step formylations of benzene, toluene, cumene,
naphthalene, fluorene, pyrene, phenanthrene, thiophene,

3468 Vy´prachticky´ and Cimrova´
Macromolecules, Vol. 35, No. 9, 2002
Sch em e 1
many electrooptical applications, such an amount of
reactive groups in the polymer is sufficient.
The π-conjugated systems (1-aminopyrene, 4-amino-
azobenzene, and 4-aminobiphenyl) were attached to
PMPSi-CHO(6 or 12 mol %) by acid-catalyzed conden-
sation of the aldehyde with amino groups yielding Schiff
bases (PMPSi-Pyr, PMPSi-Azo, PMPSi-BiPh). In the
case of PMPSi-Pyr and PMPSi-Azo almost total
conversion of aldehyde groups to the corresponding
Schiff bases was observed and proved by spectral
analysis. 1H NMR spectra of modified polymers (Figure
2) show that the azomethine hydrogens (NdCH) around
8.5 ppm are accompanied by traces of aldehyde hydro-
gens (CHO) around 9.7 ppm. From integrated intensi-
ties of these protons a conversion higher than 95% could
be calculated. If we compare the spectrum of PMPSi-
CHO (Figure 4) and the spectra of PMPSi-Pyr and
PMPSi-Azo (Figure 2), the aromatic protons of pyrene
and azobenzene structural units are clearly visible. (The
spectra of model compounds NB-Pyr and NB-Azo are
shown for convenience.) Using molar absorption coef-
ficient of NB-Pyr and NB-Azo (Figure 3), the content
of pyrene units in PMPSi-Pyr and azobenzene units
in PMPSi-Azo (Figure 5) was determined by UV-vis
spectrometry in THF. The results are in very good
agreement with the aldehyde content in PMPSi-CHO
etc., were performed41 using either dichloromethyl
methyl ether or butyl dichloromethyl ether in the yields
of 60-90%. The reaction takes place via an unstable
aryl(alkoxy)methyl chloride intermediate. On heating,
this decomposes to an aromatic aldehyde and the
corresponding alkyl chloride, with water yielding an
aromatic aldehyde, alcohol, and hydrochloric acid
(Scheme 1; Ar ) aryl, R ) methyl or butyl). This syn-
thetic method was also successfully used for formylation
of polystyrene42 where the aldehyde functionalization
of aromatic rings reached 50%.
We applied different reaction conditions for formyla-
tion of parent PMPSi-1 with dichloromethyl methyl
ether in the presence of Lewis acids (Table 1). TiCl4 was
found to be too aggressive for the silicon backbone; only
oily oligomeric products were isolated (Table 1, experi-
ments 6 and 7). When the amount of TiCl4 was de-
creased, the polymer could be isolated, but the degree
of formylation was low (Table 1, experiment 8). On the
other hand, the use of SnCl4 (a weaker Lewis acid)
appeared to be more convenient, and the reaction
proceeded smoothly. The aldehyde protons (1H) in
1
calculated from H NMR data (Table 1). Finally, FT IR
spectra of modified polymers show that the carbonyl
vibration band at 1702 cm-1 completely disappeared
during these reactions, and the new CdN vibration
band at 1620 cm-1 and other bands (see Experimental
Section) responsible for the attachment of either pyrene
or azobenzene structural units appeared. In the case of
PMPSi-BiPh the conversion of aldehyde groups was
incomplete. Integrated intensities of aldehyde (CHO)
and azomethine (NdCH) protons of 1H NMR spectrum
have shown that approximately 25% of aldehyde groups
was reacted with 4-aminobiphenyl. This estimation
could not be supported by UV-vis spectrometry because
the absorption maximum of NB-BiPh (331 nm) coin-
cides with the maximum of σ f σ* absorption band (335
nm) of PMPSi-BiPh, and only a shoulder at 360-400
nm was visible. But FT IR data of PMPSi-BiPh
supported the results received from 1H NMR. The
carbonyl vibration band at 1700 cm-1 of unreacted
aldehyde groups was detected in addition to the bands
at 1620 cm-1 (CdN) and 1170 cm-1 (aryl-N). In
contrast to formylation, the condensation reaction yield-
ing the Schiff bases proceeded without degradation of
the main chain. Molecular weight characteristics of
starting PMPSi-CHO and modified polymers were
found to be nearly unchanged (see Experimental Sec-
tion). The condensation process does not cause further
degradation of synthesized polymers in accordance with
reported results.15
1
PMPSi-CHO were clearly detected by H NMR spec-
troscopy at 9.75 ppm (Figure 4), and by comparison with
a broad signal of aromatic protons (5H) at 6.1-7.6 ppm,
the content of aldehyde groups in the polymer could be
determined (Table 1). The aldehyde C-H (2926 cm-1
)
and CdO (1702 cm-1) vibrations in addition to those of
PMPSi were detected by FT IR spectroscopy. The
aldehyde groups are assumed to be attached in the para
positions of the benzene rings for steric reasons, simi-
larly to chloromethylation.14,15 We reached the maxi-
mum degree of formylation, i.e., 12 mol % (lit.15 14-16
mol %), for a reaction time 120 min at 0 °C (Table 1,
experiment 3). The extent of formylation did not in-
crease either with increasing temperature or reaction
time (Table 1, experiments 4 and 5, respectively), in
contrast to chloromethylation.14 Approximately 50%
reduction in molecular weight was observed for PMPSi-
CHO(12 mol %) (Mw ) 21 000, Mn ) 7500) in compari-
son with parent PMPSi-1 (Mw ) 39 000, Mn ) 8600).
The formylation reaction is probably in competition with
the silicon backbone degradation (cleavage), and the
process is time-dependent. Consequently, the molecular
weight decrease for PMPSi-CHO(6 mol %) was found
to be lower (Mw ) 33 000, Mn ) 8200). This is a very
good result because in the four-step synthesis of PMPSi-
CHO substantially larger molecular weight decreases
(90%) were reported. In particular, it is the high-
molecular-weight fraction of the initial PMPSi that
seems to degrade preferentially, thereby giving rise to
a formylated polymer with lower molecular weight and
a narrower distribution. In general, the random degra-
dation cleavage of a polymer with Mw/Mn > 2 leads to
the material with polydispersity limiting to 2. In an
attempt to reach a higher degree of formylation, we
synthesized high-molecular-weight PMPSi-2 using a
phase-transfer catalyst, 18-crown-6. Even with this
material, we were not able to prepare PMPSi-CHO
containing more than 12 mol % of aldehyde groups. For
Op tica l Absor p tion a n d P h otolu m in escen ce of
Th in F ilm s. Absorption spectra of modified polysilane
films exhibited a maximum at about 330-340 nm
corresponding to the σ-conjugation on the silicon back-
bone and were similar to those measured in solutions
(cf. Figure 5). Photostability was followed from the
change of σ f σ* absorption maximum during the
illumination with a mercury lamp. The molar absorption
coefficient per silicon atom increases with molecular
weight of polysilanes,1 and consequently the chain
degradation can be evaluated by means of UV-vis
spectroscopy. Measurements were performed on several

Macromolecules, Vol. 35, No. 9, 2002
PMPSi and Its Derivatives 3469
F igu r e 6. Normalized absorbance measured at the maximum
corresponding to the σ-conjugation of the silicone backbone vs
the time of irradiation with a mercury lamp for PMPSi-1 (solid
circles) and modified polysilanes: PMPSi-Azo(6 mol %) (open
triangles), PMPSi-Azo(12 mol %) (solid triangles), PMPSi-
Pyr(6 mol %) (open circles), PMPSi-Pyr(12 mol %) (crosses),
and PMPSi-BiPh(3 mol %) (solid squares).
sets of the thin layers, and Figure 6 shows typical data.
All modified polysilanes exhibited significantly better
photostability than PMPSi. A very good stabilization
effect was obtained in polysilanes modified by the
pyrene even at low pyrene content.
PMPSi exhibits a sharp photoluminescence with high
quantum efficiency in the near-ultraviolet (NUV) region
located at 354 nm, characteristic of quasi-one-dimen-
sional excitons delocalized on the silicon backbone. It
emerges from the longest segments and gives the energy
of the singlet state for these segments.43 Phonon side-
bands are absent, indicating weak phonon coupling. The
photoluminescence study of thin films of the prepared
polymers has shown that the modification of PMPSi
quenched the PL emission peak of PMPSi at 354 nm.
An example is shown for modification with azobenzene
(Figure 7a). In comparison with PMPSi an increase in
visible PL emission intensity was observed for PMPSi-
BiPh. The PL spectrum was analyzed by multiple
Gaussian function (Figure 7b). The contribution of the
NUV emission with the band maximum at 360 nm to
the whole PL intensity is 5% only, whereas the visible
bands with maxima located at 470 and 524 nm contrib-
ute by 52% and 43%, respectively. Compared to PMPSi,
the integral PL intensity in the visible region increased
by a factor of 3. The PL spectra of PMPSi-Pyr indicate
that the excitation energy is trapped by pyrene. The
fluorescence spectrum of the pyrene consists of a
structured band of the monomer and a broad structure-
less band of excimer. The monomer fluorescence of
pyrene exhibits vibrational bands with maxima at 377,
397, and 419 nm and the maximum of the excimer band
located at about 470 nm. PL spectra of the PMPSi-Pyr
(Figure 7c) consist of the NUV emission and a broad
band in the visible region typical of the pyrene excimer.
The NUV emission could be assigned to the pyrene
monomer bands and the NUV emission of excitons
delocalized on the silicon backbone. The intensity of the
excimer emission increased with the increasing concen-
tration of the pyrene, which indicates molecular as-
sociation of pyrene in polysilane. The PL behavior in
F igu r e 7. Photoluminescence emission spectra of thin films
of (a) PMPSi (solid line), PMPSi-Azo(0.6 mol %) (dashed
line) and PMPSi-Azo(6 mol %) (dotted line); (b) PMPSi-
BiPh(3 mol %, circles), solid line is the best fit using multiple
Gaussian function with components displayed by dashed
lines; (c) PMPSi-Pyr (6 mol %, solid line) and PMPSi-Pyr
(12 mol %, dashed line). Excitation wavelength was 310 nm.

3470 Vy´prachticky´ and Cimrova´
Macromolecules, Vol. 35, No. 9, 2002
F igu r e 9. Electroluminescence spectrum (open circles) of
polymer LED: ITO/PEDT:PSS/PMPSi/Al fitted by multiple
Gaussian function (solid line) with components displayed by
dashed lines.
Ta ble 2. P a r a m eter s Obta in ed fr om th e Exp er im en ta l
Dep en d en ces ηp h vs F (See F igu r e 8) Usin g th e On sa ger
Mod el w ith th e Distr ibu tion of CT Ra d ii Given by th e
F igu r e 8. Electric field dependences of the photogeneration
efficiency (T ) 300 K) in PMPSi (open circles) and modified
polysilanes: PMPSi-Azo(6 mol %) (solid circles), PMPSi-
Pyr(6 mol %) (open squares), PMPSi-Pyr(12 mol %) (solid
squares), and PMPSi-BiPh(3 mol %) (crosses). Dashed and
solid lines were calculated using the Onsager theory with the
distribution of CT radii given by the δ- and Gaussian function,
respectively, with the parameters given in Table 2.
δ-F u n ction (r 0, η0) a n d th e Ga u ssia n F u n ction (r, η0r
)
polymer
r0 (nm)
η0
R (nm)
η0R
PMPSi
2.6
2.2
2.0
2.6
2.4
0.05
0.03
0.02
0.005
0.03
1.20
1.00
0.90
1.20
1.10
0.50
0.30
0.28
0.05
0.28
PMPSi-Pyr(6 mol %)
PMPSi-Pyr(12 mol %)
PMPSi-BiPh(3 mol %)
PMPSi-Azo(6 mol %)
modified polysilanes is a consequence of exciton mobility
in PMPSi, which was demonstrated by Tokura et al.44
They showed that defects produced by the introduction
of a bond connecting two polysilane chains efficiently
quenched the NUV emission and gave rise the visible
PL.
all the pairs are separated by the same distance r0; i.e.,
the initial spatial distribution of bound pairs is given
by the δ-function, g(r) ) (4πr2)-1δ(r-r0). The usual
Onsager model represents a first approximation only
and is valid at low electric fields (see dashed lines in
Figure 8). The full lines in Figure 8 represent the
best theoretical fits to our experimental data obtained
for the Gaussian distribution function of e-h pair radii,
g(r) ) (π-3/2R-3) exp(-r2/ R2). The parameters of the
theoretical fits are given in Table 2. The most effective
separation distance of electron-hole pairs for their
dissociation at low electric fields is close to r0, and at
high electric fields it approaches the value of R. It is
evident that the lower ηph in PMPSi-BiPh than in
PMPSi is caused by a lower value of the primary
quantum yield η0R (η0) only, whereas the separation
distances or distribution of the pair distances is the
same. In PMPSi-Pyr both shortening of the separation
distances and a decrease in the primary quantum yield
Ch a r ge Ca r r ier P h otogen er a tion . The charge
photogeneration efficiency, ηph, was measured with light
incident upon the positively charged surface and with
the intensity of light as low as possible to avoid
photodegradation. Reproducible values of ηph were
obtained, and photodegradation was not observed dur-
ing the measurements. Compared to PMPSi, in all
modified polysilanes a decrease in the ηph was observed.
The electric field dependencies of ηph are shown in
Figure 8. They were explained by the Onsager model.45
According to this model, the charge photogeneration
takes place in two steps; in the first step bound
electron-hole (e-h) pairs are formed, and in the second
step the e-h pairs dissociate into free charges. For
spherically symmetrical distribution of bound pairs, the
overall photogeneration efficiency ηph is expressed as
η
0R (η0) were observed. These results indicate a decrease
in the geminate recombination quenching at an electric
field, which could be favorable for EL. Correlation of
these results with the results of the photostability and
photoluminescence studies suggests the existence of
additional deactivation channels in the modified poly-
silanes. In PMPSi-Azo, the azobenzene isomerization
could be one of the effective channels.
η (r,F,T) ) η 4πr2f (r,F,T)g(r) dr
(3)
∫
ph
0
where η0 is the primary quantum yield, i.e., the fraction
of absorbed photons that results in e-h pairs and is
assumed to be independent of the electric field F, g(r)
an initial spatial distribution of e-h pairs, and f(r,F,T)
the Onsager’s dissociation probability45,46 of the pairs
separated by a distance r in the presence of an external
electric field F. If the usual Onsager model is assumed,
Electr olu m in escen ce. Single-layer ITO/polysilane/
Al LEDs exhibited weak white EL emission with the
highest efficiency, ηEL ∼ 0.005% photons/electrons,
detected on the ITO/PMPSi-Pyr(6 mol %)/Al devices,
which is slightly higher than that of 0.002% reported

Macromolecules, Vol. 35, No. 9, 2002
PMPSi and Its Derivatives 3471
F igu r e 11. Electroluminescence spectra of polymer LEDs:
ITO/PEDT:PSS/PMPSi-Pyr(6 mol %)/Al (solid line) and ITO/
PEDT:PSS/PMPSi-Pyr(12 mol %)/Al (dashed line).
F igu r e 10. Electroluminescence spectrum (open circles) of
polymer LED: ITO/PEDT:PSS/PMPSi-BiPh(3 mol %)/Al fitted
by multiple Gaussian function (solid line) with components
displayed by dashed lines.
for the LEDs made of pyrene single crystal.47 Our ηEL
is higher than those (ηext ∼ 0.0001%) reported in LEDs
(ITO and Al electrodes) made of PMPSi19,28 or oxygen-
cross-linked polysilanes29 and is similar to those re-
ported for LEDs (ITO and Mg:Ag electrodes) made of
silole-incorporated polysilane28 (ηext ∼ 0.001%) or poly-
[2-naphthyl(phenyl)silanediyl] doped with 2-(biphenyl-
4-yl)-5-(4-tert-butylphenyl)-1,3,4-oxadiazole20 (ηEL
0.002-0.008%). Recently, higher values of ηext ()0.1%)
were reported for LEDs (ITO and Al electrodes) made
of poly[bis(4-butylphenyl)silanediyl].25,26
∼
Double-layer devices, ITO/PEDT:PSS/polysilane/Al,
exhibited more than 1 order of magnitude higher ηEL
than those in the single-layer LEDs. The EL spectrum
of ITO/ PEDT:PSS/PMPSi/Al, shown in Figure 9, con-
sists of a NUV sharp emission peak located at 353 nm
(3.52 eV), the position of which is in good agreement
with the PL peak position, and of broad emission bands
in the visible region at 419 and 558 nm. Contrary to
Suzuki,19 who observed the sharp NUV peak at low
temperatures only, we detected such emission at room
temperature. ηEL up to 0.1% photons/electrons, deter-
mined on freshly prepared LEDs, decayed after several
runs to the value of 0.02%. As the time of LED operation
increased, the contribution of NUV emission decreased,
the contribution of visible bands increased, and the
difference in the shapes of PL and EL spectra was more
pronounced. Double-layer LEDs based on substituted
polysilanes exhibited a broad white emission. In EL
spectra of LEDs made of PMPSi-BiPh, the bands at
420 and 552 nm are dominant (Figure 10). EL spectra
of LEDs made of PMPSi-Pyr consisted of broad bands,
which could be assigned to the emission of the pyrene
excimers with a different conformation and to the
recombination centers of the structural origin. The band
at 590 nm dominates (Figure 11). The EL spectra differ
from PL ones due to different excitation, which indicates
that different recombination centers are dominant in
PL and EL. In EL, the charge trapping in the radiative
recombination centers plays an important role. ηEL in
double-layer LEDs made of PMPSi-BiPh or PMPSi-
Pyr were in the range of 0.03-0.06% photons/electrons.
F igu r e 12. Dependence of the current and spectrally inte-
grated EL intensity on the applied electric field measured on
polymer light-emitting devices: ITO/PEDT:PSS/PMPSi-BiPh-
(3 mol %)/Al (crosses), ITO/PEDT:PSS/PMPSi-Pyr(6 mol %)/
Al (open circles), and ITO/PEDT:PSS/PMPSi-Pyr(12 mol %)/
Al (solid circles).
In PMPSi-Pyr devices, higher values of ηEL were
measured for a lower pyrene content. The measured ηEL
values are in the same range as the values reported for
LEDs based on π-conjugated polymer such as poly[2,5-
bis(isopentyloxy)-1,4-phenylene]48 and are also compa-
rable with the EL efficiency measured in double-layer
devices composed of hole-transporting (triphenylamine)
and light-emitting (random copolymer of 9,9-dihexyl-
fluorene and anthracene) layers.49 Similar EL efficien-
cies were reported for the devices based on PMPSi
containing anthracene units.32
The onset of the EL emission was at ca. 1.5 × 108 V
m-1, being connected with a current increase at high
electric fields (see Figure 12). The hole mobility of
PMPSi (10-8 m2 V-1 s-1) is not significantly modified
by low substitution and the barrier height at the ITO/

3472 Vy´prachticky´ and Cimrova´
Macromolecules, Vol. 35, No. 9, 2002
polysilane interface is small, so the polysilane LEDs are
typical hole-only devices. The ionization potential (IP),
which corresponds to the HOMO energy level, was
determined from the cyclovoltammetric measurements.
In PMPSi the onset of oxidation potential (Eox) was 0.7
V relative to the Ag/AgCl reference electrode. The Eox
of modified polysilanes (Eox ∼ 0.72-0.78 V) was slightly
shifted to higher voltages compared with PMPSi. Simi-
lar to PMPSi and other polysilanes,9 the modified
polysilanes oxidized irreversibly. Using an empirical
relationship,50 IP ) 4.4 + Eox (eV), the IP values of 5.1-
5.2 eV were estimated. The ITO work function usually
varies from 4.7 to 4.9 eV, so the expected height of the
injection barrier for holes is lower than 0.5 eV and could
further be reduced by using PEDT:PSS. On the other
hand, the injection barrier for electrons in polysilanes
is several times higher. However, significant hole trap-
ping at the radiative centers near the polysilane/Al
interface can improve the electron injection due to the
space charges produced by trapped holes. The analysis
of the current characteristics shows the influence of
the space charges. Two regimes can be distinguished.
At lower electric fields (F), the quadratic dependence
of the current on F is characteristic of the space-
charge-limited current (SCLC) behavior.51 At higher F,
when EL emission starts, the current I increases more
steeply, I ∼ Fm, with m at about 9-10, which can be
explained in the frame of SCLC with an exponential
trap distribution.51 While the current is dominated by
holes, the EL intensity IEL is determined by electron
supply. The dependence of IEL on F shows a Fowler-
Nordheim behavior due to tunneling of minority carriers
(electrons).52
reduction caused by formylation is less (50%) than that
in the multistep procedure (90%). The attachment of
aromatic amines yielding Schiff bases proceeded smoothly
and without further degradation of the main chain. For
PMPSi-Pyr and PMPSi-Azo, almost total conversion
of aldehyde groups to the corresponding Schiff bases was
observed, but in the case of PMPSi-BiPh, the aldehyde
groups conversion was only 25%.
Photophysical properties (photostability, photolumi-
nescence, and charge photogeneration) of modified poly-
silanes were studied, and polymer light-emitting devices
were fabricated and characterized. In comparison with
PMPSi, quenching of NUV emission and an increase in
the visible PL emission intensity were observed in
modified polysilanes. LEDs emitted white light, and the
electroluminescence spectra differed from photolumi-
nescence spectra. Charge trapping in the radiative
recombination centers plays an important role in EL
emission. The EL efficiency up to 0.06% photon/electron
was detected in double-layer LEDs, which is 1 order of
magnitude higher than that in single-layer devices.
Chemical substitution improved the photostability of
PMPSi; therefore, the modified polysilanes are promis-
ing for use in polymer-blend LEDs to modify transport
and injection properties and/or to prevent aggregation
of the conjugated polymers. The Onsager model with
the initial distribution of separation distances of elec-
tron-hole pairs given by the Gaussian function is a
suitable model for description of the electron-hole pair
dissociation during the photogeneration process in poly-
silanes. The most effective separation distances of
electron-hole pairs in the charge photogeneration
process were determined to be 2.0-2.6 and 0.9-1.2 nm
at low and high electric fields, respectively.
As follows from the discussion above, the limiting
factor in polysilane LEDs is electron injection. There-
fore, better device performances are expected using
metals with low work functions such as Ca or by
introducing an electron-transporting layer. The EL
efficiencies of polysilane LEDs are low; therefore, these
materials are particularly intended for the use in
multilayer or blend devices as hole transporting poly-
mers to modify transport and injection properties.
Recently, we have shown an improved performance (an
increase in ηEL up to 25 times and improved stability)
of blend LEDs made of an EL polymer (substituted poly-
(terphenyl-4,4′′-diylvinylene)) and a hole-transporting
polymer (poly[biphenyl-4-yl(phenyl)silanediyl])53 com-
pared to the device made of neat EL polymer. Further,
we achieved an increase in ηEL (up to 30 times) and
improvement of the stability of blue emission in LEDs
made of blends composed of poly[2,5-bis(isopentyloxy)-
1,4-phenylene] (SPPP) and PMPSi, compared to the case
of SPPP LEDs.33 The results of our present research
showed that blending of PMPSi-Pyr with poly(9,9-
dihexadecylfluorene-2,7-diyl) (SPF) leads to an increase
in ηEL (ca. 2 orders of magnitude) and stability, com-
pared to the case of SPF LEDs.
Ack n ow led gm en t. We thank the Grant Agency of
the Czech Republic for supporting this work by Grant
102/98/0696 and the Volkswagen Foundation.
Refer en ces a n d Notes
(1) Miller, R. D.; Michl, J . Chem. Rev. 1989, 89, 1359.
(2) Watanabe, A.; Miike, H.; Tsitsumi, Y.; Matsuda, M. Macro-
molecules 1993, 26, 2111.
(3) Trujilo, R. E. J . Organomet. Chem. 1980, 198, C27.
(4) West, R.; David, L. D.; Djurovich, P. I.; Stearley, K. L.;
Srinivasan, K. S. V.; Yu, H. J . Am. Chem. Soc. 1981, 103,
7352.
(5) Weson, J . P.; Williams, T. C. J . Polym. Sci., Polym. Chem.
Ed. 1980, 18, 959.
(6) J edlinski, Z. J .; Kurcok, P.; Nozirow, F. Macromol. Rapid
Commun. 1997, 18, 483.
(7) J ones, R. G.; Benfield, R. E.; Evans, P. J .; Swain, A. C. J .
Chem. Soc., Chem. Commun. 1995, 1465.
(8) Ho Yi, S.; Maeda, N.; Suzuki, T.; Sato, H. Polym. J . 1992,
24, 865.
(9) Cleij, T. J .; King, J . K.; J enneskens, L. W. Macromolecules
2000, 33, 89.
(10) Harrah, L. A.; Zeigler, J . M. Macromolecules 1987, 20, 2037.
(11) Horiguchi, R.; Onishi, Y.; Hayase, S. Macromolecules 1988,
21, 304.
(12) Stu¨ger, H.; West, R. Macromolecules 1985, 18, 2349.
(13) Ban, H.; Sukegawa, K.; Tagawa, S. Macromolecules 1987, 20,
1775.
Con clu sion s
(14) J ones, R. G.; Benfield, R. E.; Swain, A. C.; Webb, S. J .; Went,
We have developed a one-step synthesis of poly-
[(formylphenyl)methylsilanediyl], which was utilized as
the functionalized polymer precursor for the attachment
of π-conjugated systems to the polysilane backbone. 1H
NMR, UV-vis, and FT IR spectroscopies proved the
presence of the aldehyde functionality. We reached the
maximum content of reactive aldehyde groups in the
functionalized polymer, 12 mol %. The molecular weight
M. J . Polymer 1995, 36, 393.
(15) Km´ınek, I.; Brynda, E.; Schnabel, W. Eur. Polym. J . 1991,
27, 1073.
(16) Cimrova´, V.; Km´ınek, I.; Nesˇpu˚rek, S.; Schnabel, W. Synth.
Met. 1994, 64, 271.
(17) Kido, J .; Nagai, K.; Okamoto, Y.; Skotheim, T. Appl. Phys.
Lett. 1991, 59, 2760.
(18) Suzuki, H.; Meyer, H.; Simmerer, J .; Yang, J .; Haarer, D.
Adv. Mater. 1993, 5, 743.

Macromolecules, Vol. 35, No. 9, 2002
PMPSi and Its Derivatives 3473
(19) Suzuki, H. Adv. Mater. 1996, 8, 657.
(36) Tabei, K.; Saitou, E. Bull. Chem. Soc. J pn. 1969, 42, 2693.
(37) Greenham, N. C.; Friend, R. H.; Bradley, D. C. C. Adv. Mater.
1994, 6, 491.
(38) Chen, I.; Mort, J .; Tabak, J . H. IEEE Trans. Electron Devices
1972, 19, 413.
(39) Ayres, J . T.; Mann, C. K. J . Polym. Sci., Polym. Lett. Ed. 1965,
3, 505.
(40) Dilke, M. H.; Eley, D. D. J . Chem. Soc. 1949, 2613.
(41) Rieche, A.; Gross, H.; Ho¨ft, E. Chem. Ber. 1960, 93, 88.
(42) Kahovec, J . Polym. Bull. (Berlin) 1981, 4, 731.
(43) Elsner, A.; Mahrt, R. F.; Pautmeier, L.; Baessler, H.; Stolka,
M. Chem. Phys. 1991, 150, 81.
(44) Tokura, H.; Tachibana, H.; Kishida, H.; Morimoto, Y.;
Matsumoto, M. J . Phys. Soc. J pn. 1994, 63 (Suppl. B), 30.
(45) Onsager, L. Phys. Rev. 1938, 54, 554.
(46) Mozumder, A. J . Chem. Phys. 1974, 60, 4300.
(47) Gonzales-Basurto, J .; Burshtein, Z. Mol. Cryst. Liq. Cryst.
1975, 31, 211.
(48) Remmers, M.; Neher, D.; Gru¨ner, J .; Friend, R. H.; Gelinck,
G. H.; Warman, J . M.; Quattrocchi, C.; dos Santos, D. A.;
Bredas, J . L. Macromolecules 1996, 29, 7432.
(49) Chen, J . P.; Markiewicz, D.; Lee, V. Y.; Kla¨rner, G.; Miller,
R. D.; Scott, J . C. Synth. Met. 1999, 107, 203.
(50) Bre´das, J .-L.; Silbey, R.; Boudreaux, D. S.; Chance, R. R. J .
Am. Chem. Soc. 1983, 105, 6555.
(51) Kao, K. C.; Hwang, W. Electrical Transport in Solids;
Pergamon Press: Oxford, 1981; pp 151-169.
(52) Berleb, S.; Bru¨tting, W.; Schwoerer, M.; Wehrman, R.;
Elschner, A. J . Appl. Phys. 1998, 83, 4403.
(20) Fujii, A.; Yoshida, M.; Yoshimoto, K.; Ohmori, Y.; Yoshino,
K.; Ueno, H.; Kakimoto, M.; Kojima, H. Synth. Met. 1997,
85, 1261.
(21) Hoshino, S.; Suzuki, H.; Fujiki, M.; Morita, M.; Matsumoto,
N. Synth. Met. 1997, 89, 221.
(22) Hoshino, S.; Suzuki, H.; Fujiki, M.; Morita, M.; Matsumoto,
N. Mol. Cryst. Liq. Cryst. 1998, 315, 205.
(23) Suzuki, H.; Hoshino, S.; Yuan, C.-H.; Fujiki, M.; Toyoda, S.;
Matsumoto, N. Thin Solid Films 1998, 331, 64.
(24) Furukawa, S. Thin Solid Films 1998, 331, 222.
(25) Hoshino, S.; Furukawa, K.; Ebata, K.; Breyl, I.; Suzuki, H.
J . Appl. Phys. 2000, 88, 3408.
(26) Suzuki, H.; Hoshino, S.; Furukawa, K.; Ebata, K.; Yuan, C.-
H.; Bleyl, I. Polym. Adv. Technol. 2000, 11, 460.
(27) Hoshino, S.; Ebata, K.; Furukawa, K. J . Appl. Phys. 2000,
87, 1968.
(28) Adachi, A.; Yasuda, H.; Sanji, T.; Sakurai, H.; Okita, K. J .
Lumin. 2000, 87-89, 1174.
(29) Hiraoka, T.; Majima, Y.; Murai, S.; Nakano, Y.; Hayase, S.
Polym. Adv. Technol. 1997, 8, 465.
(30) Malliaris, G. G.; Herrema, J . K.; Wildeman, J .; Wieringa, R.
H.; Gill, R. E.; Lampoura, S. S.; Hadziioannou,G. Adv. Mater.
1993, 5, 721.
(31) Yoshino, K.; Fujii, A.; Nakayama, H.; Lee, S.; Naka, A.;
Ishikawa, M. J . Appl. Phys. 1999, 85, 414.
(32) Satoh, S.; Suzuki, H.; Kimata, Y.; Kuriyama, A. Synth. Met.
1996, 79, 97.
(33) Cimrova´, V.; Vy´prachticky´, D.; Pecka, J .; Kotva, R. SPIE Proc.
Vol. 3939; Organic Photonic Materials and Devices II, 24-
26 J an 2000, San J ose, CA.
(53) Cimrova´, V.; Neher, D.; Remmers, M.; Kmı´nek, I. Adv. Mater.
1998, 10, 676.
(34) Ricketts, J . A.; Cho, C. S. J . Org. Chem. 1961, 26, 2125.
(35) Iqbal, A. F. M. J . Org. Chem. 1972, 37, 2791.
MA010736W